(See BRACIL for details)
My first degree was in Business Administration, majoring in Finance (Chinese University of Hong Kong). After a few years' work in industry, I studied for my Master degree in Computer Science at University of Essex. Inspired by Professor Jim Doran, I stayed on to do a PhD in artificial intelligence.
I have broad interest in applied artificial intelligence, especially constraint satisfaction, financial applications, planning, temporal reasoning and scheduling. My major interest are constraint satisfaction and optimization and computational finance, economics and management.
GENET is a connectionist approach to constraint satisfaction. The idea to take advantage of massive parallel hardware in solving constraint satisfaction problems. This work has been extended to Guided Local Search (GLS), which is a meta-hill climbing strategy which borrowed its ideas from Operations Research. GLS has been applied to a non-trivial number of problems, including some real life problems.
In the ACS project, which has just finished, we develope strategies which dynamically allocate algorithms to problems depending on measurable problem features, and monitor the progress of algorithms while searching (switching to other algorithms when needed).
The Computer-aided Constraint-Programming project is concerned about the software engineering aspect of constraint satisfaction. It aims to bring constraint technology to non-expert users. Therefore, it is concerned with computer-aided problem formulation, algorithm selection. The Human Computer Interaction is also looked at in this project.
In computational finance, economics and management, I lead the Computational Finance Research Laboratory at University of Essex. I am also the Deputy Director of the Centre for Computational Finance and Economic Agents, an interdisciplinary research centre at University of Essex.
In computational economics, I lead a project on automated bargaining. Bargaining is a research topic that has produced no fewer than five Nobel Prize laureates. Perfect rationality has been the basis of most theoretical results. Once perfect rationality is relaxed, many theoretical results will have to be re-examined. We use co-evolutionary using genetic programming to approximate Nash equilibrium. Co-evolution takes a local improvement approach -- in the form of arm race -- which does not rely on the perfect rationality assumption. Our work complements theoretical approach. Comparing and constrasting results of both approaches should bring the bargaining field forward.
In management, we have embarked on a project on staff empowerment, a modern management concept which aims to boost morale. This project involves many technical issues. Automated bargaining and heuristic search play an important part in our market-based approach to staff empowerment. Multi-objective optimization and machine learning are part of this research. Research involves designing business processes in order to assess risk and explore business opportunities. The computational techniques to be developed is general, though BTís scheduling problem will be used as a test-bed.
I have worked with many companies and institutes, include GEC Marconi Research Centre, British Telecom, The Commonwealth Secretariat, and some finance companies.